Electricity/Solar Update

A look at the latest Jedlovec House eletricity usage and solar production from PPL and Enphase.

Load packages

library(tidyverse)
library(lubridate)
library(hms)
library(readxl)

Load PPL data

Transform PPL Data

hourly_ppl_pivot <- ppl_15mins %>% 
  rename(date = Date) %>% 
  pivot_longer(!c("Account Number", "Meter Number", date, "Read Type", Min, Max, Total), names_to = "time", values_to = "kWh") 

rename(ppl_15mins, date = Date)

hourly_ppl_pivot <- hourly_ppl_pivot %>% 
  mutate(time = parse_time(time, '%H:%M %p'), month = month(date, label=TRUE), year = year(date), yday = yday(date), wday = wday(date, label=TRUE))

(hourly_ppl_net <- hourly_ppl_pivot %>% 
  filter(`Read Type` == "kWh Net"))

Load Enphase data

import <- read_csv("enphase_history_230701.csv")

Rows: 42432 Columns: 2── Column specification ────────────────────────────────────────────────────────────────────────────────────────
Delimiter: ","
chr (1): Date/Time
dbl (1): Energy Produced (Wh)
ℹ Use `spec()` to retrieve the full column specification for this data.
ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.

update <- read_csv("hourly_generation_230625_240510.csv")

Rows: 30816 Columns: 2── Column specification ────────────────────────────────────────────────────────────────────────────────────────
Delimiter: ","
chr (1): Date/Time
dbl (1): Energy Produced (Wh)
ℹ Use `spec()` to retrieve the full column specification for this data.
ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.

update <- update %>% 
  mutate(`Date/Time` = as.POSIXct(`Date/Time`,format="%m/%d/%Y %H:%M",tz=Sys.timezone())) %>% 
  filter(`Date/Time` >= as.POSIXct('07/01/2023 00:00',format="%m/%d/%Y %H:%M",tz=Sys.timezone()) )

import <- import %>% 
  mutate(`Date/Time` = as.POSIXct(`Date/Time`,format="%m/%d/%Y %H:%M",tz=Sys.timezone())) 

full_hist <- rbind(import,update)

full_hist %>% arrange(desc(`Date/Time`))

(hourly_production <- full_hist %>% 
  rename(datetime = `Date/Time`, energy_produced_Wh = `Energy Produced (Wh)`) %>% 
  mutate(datetime = as.POSIXct(datetime,format="%m/%d/%Y %H:%M",tz=Sys.timezone())) %>% 
  mutate(date = date(datetime), time = as.hms(format(datetime, format = "%H:%M:%S")), month = month(datetime, label=TRUE), year = year(datetime), day = day(datetime), yday = yday(datetime), monthday = format(datetime, "%m-%d"), wday = wday(datetime, label=TRUE), equinox_day = (yday + 10) %% 365, equinox_group = floor((equinox_day+15)/30)*30)
)

NA

Spot-check Enphase Should see lifetime production by day and by hour

ggplot(hourly_production, aes(datetime, energy_produced_Wh)) +
  geom_point()

ggplot(hourly_production, aes(time, energy_produced_Wh)) +
  theme(axis.text.x = element_text(angle = 90)) +
  geom_point()

Net + Produced = Consumed

Calculate production for first year of solar panels, second year, etc.

hourly_electricity <- hourly_electricity %>% 
  mutate(solar_year = factor(case_when(
                        datetime < as.POSIXct('04/16/2022 00:00',format="%m/%d/%Y %H:%M",tz=Sys.timezone()) ~ 0,
                        datetime >= as.POSIXct('04/16/2022 00:00',format="%m/%d/%Y %H:%M",tz=Sys.timezone())  &
                          datetime < as.POSIXct('04/16/2023 00:00',format="%m/%d/%Y %H:%M",tz=Sys.timezone()) ~ 1,
                        datetime >= as.POSIXct('04/16/2023 00:00',format="%m/%d/%Y %H:%M",tz=Sys.timezone())  &
                          datetime < as.POSIXct('04/16/2024 00:00',format="%m/%d/%Y %H:%M",tz=Sys.timezone()) ~ 2,
                        TRUE ~ 3)
         )) %>% 
  mutate(yday = yday(date))

hourly_electricity %>% 
  group_by(solar_year) %>% 
  summarize(net_kWh = sum(net_kWh), produced_kWh = sum(produced_kWh), consumed_kWh = sum(consumed_kWh))

NA

Interesting! Produced kWh went down by 5%, but consumed kWh went down by 10% despite the fact that we got an electric car! Let’s explore that further.

daily_electricity <- hourly_electricity %>% 
  group_by(solar_year, yday) %>% 
  summarize(net_kWh = sum(net_kWh), produced_kWh = sum(produced_kWh), consumed_kWh = sum(consumed_kWh))

`summarise()` has grouped output by 'solar_year'. You can override using the `.groups` argument.

ggplot(daily_electricity, aes(yday, consumed_kWh, group=solar_year, color=solar_year)) + 
  geom_point() +
  geom_smooth(span=0.3) +
  scale_x_continuous(breaks = c(1, 32, 60, 91, 121, 152, 182, 213, 244, 274, 305, 335), labels = month.abb) +
  theme(axis.text.x = element_text(angle = 90)) 

Look how our electricity consumption got much less predictable, more uneven, after we purchased an EV and installed a level 2 charger in late January/early February 2023!

We probably need to filter out the EV charging and compare that separately.

hourly_electricity %>% summarize(min_date = min(date), max_date = max(date))

electricity_by_time <- hourly_electricity %>% 
  filter(solar_year == 1 | solar_year == 2) %>% 
  #filter(date <= as.POSIXct('12/31/2022 00:00',format="%m/%d/%Y %H:%M",tz=Sys.timezone()) ) %>% 
  group_by(time, solar_year) %>% 
  summarize(produced_kWh = mean(produced_kWh), consumed_kWh = mean(consumed_kWh), net_kWh = mean(net_kWh)) %>% 
  arrange(time)

`summarise()` has grouped output by 'time'. You can override using the `.groups` argument.

electricity_by_time

ggplot(electricity_by_time, aes(x=time, y=consumed_kWh, group=solar_year, color=solar_year)) +
  geom_point() 

  #labs(colour="",x="Time of Day",y="Electricity Consumption (kWh)")+
  #scale_color_manual(values = c("red","black","green")) +
  #ggtitle("Home Electricity in 15-Minute Intervals (Since April 15, 2022)")

Wow, look how much less was used during the day in year 2!!! Like a 50% reduction!

Is this due to milder weather?

Let’s look at it by month:

#electricity_by_month <- 
hourly_electricity %>% 
  mutate(month = month(date)) %>% 
  group_by(solar_year, month) %>% 
  summarize(consumed_kWh = mean(consumed_kWh)) %>% 
  arrange(month, solar_year) %>% 
  pivot_wider(names_from = month, values_from = consumed_kWh) %>% 
  arrange(solar_year)

`summarise()` has grouped output by 'solar_year'. You can override using the `.groups` argument.

#Look at limited window during day (no EV charging)
 
# hourly_electricity %>% 
#   filter(time == '12:00:00') %>% 
#   #filter(time > as.POSIXct('04:00:00',format="%H:%M:$s",tz=Sys.timezone()) ) %>%  #& time < as.POSIXct('18:00',format="%H:%M",tz=Sys.timezone())) %>% 
#   mutate(month = month(date)) %>% 
#   group_by(solar_year, month) %>% 
#   summarize(consumed_kWh = mean(consumed_kWh)) %>% 
#   arrange(month, solar_year) %>% 
#   pivot_wider(names_from = month, values_from = consumed_kWh) %>% 
#   arrange(solar_year)

Look at specific months

electricity_by_month_time <- hourly_electricity %>% 
  mutate(month = as_factor(month(date))) %>% 
  filter(solar_year == 1 | solar_year == 2) %>% 
  filter(month == 10) %>% 
  group_by(solar_year, month, time) %>% 
  summarize(produced_kWh = mean(produced_kWh), consumed_kWh = mean(consumed_kWh), net_kWh = mean(net_kWh)) %>% 
  arrange(solar_year, month, time)

`summarise()` has grouped output by 'solar_year', 'month'. You can override using the `.groups` argument.

ggplot(electricity_by_month_time, aes(time, consumed_kWh, group=solar_year, color=solar_year)) +
  #facet_grid(rows = vars(month)) +
  geom_point(size=0.5) +
  ggtitle("Home Electricity in 15-Minute Intervals") +
  ylab("Net Electricity Consumption (kWh)")

Example day with EV charging

sample_day <- hourly_electricity %>% 
  filter(solar_year == 1 | solar_year == 2) %>% 
  filter(yday == 157) %>% 
  group_by(date, solar_year, yday, time) %>% 
  summarize(produced_kWh = mean(produced_kWh), consumed_kWh = mean(consumed_kWh), net_kWh = mean(net_kWh)) %>% 
  arrange(date, solar_year, yday, time)

`summarise()` has grouped output by 'date', 'solar_year', 'yday'. You can override using the `.groups` argument.

ggplot(sample_day, aes(time, consumed_kWh, group=solar_year, color=solar_year)) +
  #facet_grid(rows = vars(month)) +
  geom_point(size=0.5) +
  ggtitle("Home Electricity in 15-Minute Intervals") +
  ylab("Electricity Consumption (kWh)")

Ok, let’s detect and remove EV charging sessions.

bcp

$x
 [1] -0.4717343  0.7711507 -1.5990938 -0.7508534  0.8046654 -0.6484233  0.1813083  0.7249717 -0.6922743
[10]  1.0614834  3.6704000  4.0869019  4.0584831  4.9843196  5.1803859  5.1696722  5.0958858  5.5062540
[19]  5.3078164  5.3926576

$max_p
           [,1] [,2] [,3]
 [1,] 0.9009547    0    0
 [2,] 0.8015196    0    1
 [3,] 0.7616943    0    2
 [4,] 0.7018043    0    3
 [5,] 0.6985363    0    4
 [6,] 0.6588125    0    5
 [7,] 0.5838624    0    6
 [8,] 0.6049459    0    7
 [9,] 0.5252828    0    8
[10,] 0.3619282    0    9
[11,] 0.4815652    0   10
[12,] 0.5369016   10    0
[13,] 0.9141713   10    1
[14,] 0.9182526   10    2
[15,] 0.9211320   10    3
[16,] 0.9242694   10    4
[17,] 0.9249073   10    5
[18,] 0.9268006   10    6
[19,] 0.9281916   10    7
[20,] 0.0000000    0    0

$parameters
theta alpha  beta th_cp 
  0.9   1.0   1.0   0.5 

$series_length
[1] 20

$result
NULL

attr(,"class")
[1] "BayesCP"